A Staff Engineer's Guide to Inventing Work

Why we picked it · the editor's summary

Sujith Jay's claim is that an engineering-led platform team rarely fails with an empty backlog; it fails with one assembled from the loudest signals, usually the crash and sometimes the skip-level. The guide sorts the alternatives into four sources: systems (postmortems, cost, toil), users (regular conversations about recent tasks, and workarounds treated as prototypes), the organization (OKRs, what managers keep repeating, debris from past migrations), and the industry, followed on purpose with a lag. One practice stands out: writing design documents for systems that already exist, to find decisions that have quietly expired. We'd start there; it is the one that turns up work nobody filed.